Once reactivated, previously consolidated memories destabilize and have to be reconsolidated to persist, a process that might be altered non-invasively by interfering learning immediately after reactivation. Here, we investigated the influence of interference on brain correlates of reactivated episodic memories for emotional and neutral scenes using event-related potentials (ERPs). To selectively target emotional memories we applied a new reactivation method: rapid serial visual presentation (RSVP). RSVP leads to enhanced implicit processing (pop out) of the most salient memories making them vulnerable to disruption. In line, interference after reactivation of previously encoded pictures disrupted recollection particularly for emotional events. Furthermore, memory impairments were reflected in a reduced centro-parietal ERP old/new difference during retrieval of emotional pictures. These results provide neural evidence that emotional episodic memories in humans can be selectively altered through behavioral interference after reactivation, a finding with further clinical implications for the treatment of anxiety disorders.
Changing unpleasant or even traumatic memories is one of the major challenges for clinical interventions (Parsons & Ressler, 2013). It is a well-established finding that the formation of emotional long-term memories is mediated by the adrenergic system and depends on the interaction between amygdala and hippocampus (McGaugh, 2000). Following retrieval consolidated memories return to an unstable state and have to be stabilized again into a persisting memory, a process that is known as memory reconsolidation (Nader, Schafe, & LeDoux, 2000). Recent animal and human research has successfully targeted the modification of conditioned fear memories through blockade of reconsolidation after reactivation of these memories by pharmacological agents, such as beta-adrenergic receptor blockers (Nader and Hardt, 2009 and Nader et al., 2000). To date, psychophysiological and emerging behavioral evidence exists for successful blockade of reconsolidation in humans using pharmacological or behavioral interventions (Chan and LaPaglia, 2013 and Schiller et al., 2010; for review see Agren, 2014). For instance, it has been shown that new learning after reactivation of previously learned material may impair (Wichert, Wolf, & Schwabe, 2011) or update (Hupbach, Gomez, Hardt, & Nadel, 2007) memory. Despite promising behavioral evidence only few brain imaging studies (Agren et al., 2012; for review see Schwabe, Nader, & Pruessner, 2014) investigated human brain function underlying altered memory representations following reconsolidation blockade in humans. Agren et al. (2012) found that behavioral disruption of fear reconsolidation significantly decreased memory trace activity in the amygdala. Using pharmacological reconsolidation blockade, Schwabe, Nader, Wolf, Beaudry, and Pruessner (2012) found that impairments in memory for emotional materials were associated with altered amygdala and hippocampus activation.
In the present study we used event-related potentials (ERPs) to investigate the brain dynamics underlying episodic emotional memories. ERPs provide non-invasive measures of neural activity with high time resolution (ms range) and are thus well suited to examine the neural networks underlying human memory (Voss & Paller, 2008). In recognition memory tasks it is a key finding that ERPs during the retrieval of previously encoded “old” items evoke more positive going waveforms than correctly classified “new” items (Rugg et al., 1998). This so-called ERP old/new effect is most prominent over centro-parietal brain sites, starting at about 500 ms post-stimulus, and has been associated with hippocampus-dependent explicit recollection (Düzel, Vargha-Khadem, Heinze, & Mishkin, 2001). Numerous studies have found that the late ERP old/new effect is specifically enhanced for emotionally arousing compared to neutral stimuli (Weymar & Hamm, 2013), an effect that can be abolished by pre-encoding beta-adrenergic blockade (Weymar et al., 2010). The beta-adrenergic system also mediates the reconsolidation of conditioned fear memory in humans (Kindt, Soeter, & Vervliet, 2009).
Here, we tested whether encoding of new information after reactivation of previously encoded memories interferes with the reconsolidation of reactivated emotional and neutral episodic memories and whether this interference can be traced in the neural signature of episodic memory. We applied a new method for memory reactivation – rapid serial visual presentation (RSVP). During this task all 90 previously encoded pictures were presented in a rapid stream so that the entire reactivation lasted only for 30 s (Versace, Bradley, & Lang, 2010). Based on previous research four experimental groups were included in the design (see Fig. 1). In two groups old memories were reactivated using the 30 s RSVP after one day and reconsolidation of these old memories were interrupted after 10 min by an interfering task (encoding of new emotional and neutral scenes) in one group but not in the other. Two groups without the reactivation manipulation were added as control groups. One of these groups just received the interference task on day two to assess the unique influence of the interference task.
Full-size image (30 K)
Fig. 1.
Experimental design. The four experimental groups (Reactivation, Reactivation + interference, Non-Reactivation Interference, Non-Reactivation Control) and time intervals between the experimental sessions (Encoding, Reactivation, Interference learning task and Recognition).
Figure options
We expected that new learning, compared to no learning, would impact reconsolidation of previously reactivated pictures. Because pictures were reactivated through RSVP – favoring the processing of salient stimuli – we predicted that particularly emotional memories would be affected by the interference task resulting in impaired recognition memory performance and smaller centro-parietal old/new difference in the ERPs.
